Do you Sleep, Hibernate, or Shut Down Your computer?

Community
41

Hard drives tend to go first. An interaction of high speed, delicate moving parts and the expansion/contraction of metal due to heat.

Cheap power supplies go next. They age quickly compared to other hardware because they’re subjected to your home power system (overvoltage, undervoltage, spikes) and because of heat. They produce a lot of heat of their own, and have to contend with all the heat the rest of the computer is producing. Lots of electrical parts have shorter life spans as you subject them to higher operating temperatures.

Pretty much everything else in a computer should last a good 10 years at least. Occasionally an optical drive might have a problem, but I’ve never actually had to replace one because it stopped functioning. At a lower level, well, I’m not an electrical engineer, but I would guess capacitors are most at risk. Overvolt them or get 'em too hot and they’ll bubble and eventually explode through their cap. In a PSU, the risk is from overvoltages and heat, and on the motherboard, just from heat assuming the PSU does its job.

I read a story not too many months ago about a rash of electronics dying after 1-2 years, across all kinds of markets from DVD players to cable boxes and computer monitors. There were some Chinese companies that stole an incomplete formula for the fluid in electrolytic capacitors and were producing and selling them anyway. They turned out to have a much shorter lifespan than typical, and since the producers didn’t know, much shorter than rated as well. I sometimes wonder if one of my monitors has those bad capacitors in it. I can’t turn the monitor off, because turning it on is a 45 minute ordeal of getting capacitors to charge up enough to get the backlight lit. I can hear the high pitch whine.

42

So it sounds like for the typical “away” time (i.e. 8 - 16 hours), the best thing (to extend lifetime, not conserve energy) would be a compromise between full on and full off. Shutting down those parts that are more heat sensitive and keeping up those that take a bigger hit for start-up. I guess I should look more closely at exactly what sleep and hibernate do.

43

Again a popular myth without numbers. Why are semiconductors thermal cycled so many hundreds of degrees during construction - without damage? Tens of degree temperature variations are only destructive when observation - not tempered by knowledge or numbers - results in junk science conclusions.

Back in junior high, when I built my first OP amp design, I was also concerned that the OP was so hot. So I read the specs. It was not too hot. Later, old engineers provided perspective. If you touch it and don’t leave skin, then it is cool enough.

Reality - heat causes timing and threshold changes. That temporary change causes crashes - not hardware damage. To have hardware damage means it must be so hot as to leave skin.

Does heat reduce life expectancy? Yes. Then we apply numbers. Does that sentence sound familiar? Instead of lasting 400 years, it will now last maybe 200 years.

A perfectly ideal temperature for any properly constructed computer is a room at 100 degrees F. Any computer that does not work at those temperatures has some other defect. Temperature is traditionally an ideal diagnostic tool. Those who never learn this stuff will blame temperature rather than fix the problem.

How to find defective hardware before it fails months later? Execute diagnostics in a 100+ degree room. If it fails, then replace the completely defective semiconductor / component. Heat is a tool to locate defective hardware. Others who never learn these concepts will cure symptoms - ie more fans.

Tim Allen mocked those who use such reasoning - “More Power”.

What is the most common reason for failure? Manufacturing defects. Dan Grossman’s electrolytic failure is a perfect example of manufacturing defects appearing months or years later. How many instead blamed that failure on heat before the problem was publicized?

Another example of people using observation to make junk science conclusions: He powered off the machine. Then it would not power on. He therefore blamed power cycling. The analysis: A pullup resistor had only one function. To bootstrap the power supply controller during power up. That resistor is constantly powered and failed sometime during normal operation. A manufacturing defect. Normal operation (not power cycling) caused the failure. Observation without learning facts resulted in another junk science conclusion.

Blaming heat for failure is a classic example of junk science reasoning so common when most all computer techs do not even know how electricity works. When he forgets what is required to have a fact. When observation is the only reason for a conclusion: junk science.

A standard computer has more than sufficient cooling with one 80 mm chassis fan. That is contrary to popular myths, in part, because a majority do not first do simple thermal equations or read numbers from datasheets. That need for cooling: A defective computer failed when it was warm. And a defective computer worked with more fans. Another junk science concluson based only upon observation.

44

I think someone has misinterpreted what I was claiming.

I was arguing for ALWAYS turning off lights when not required since for most even a second or so of continuous use uses more power than the additional power used when turning them on and all modern lights use less than about 10 seconds worth of extra power in turning them on.

As for the bulbs blowing sooner due to turning them on and off more often it was someone else who first made that claim. My only comments were that there were some lights that had been on constantly for over 100 years and that the risk involved in getting to at least one of them to replace it when it does blow was a good reason why it was arranged to hopefully last as long as possible before that happens.

I was informed that was junk science and that the surge when turning lights on has no effect on their effective life. To that I responded that every light that I have seen have the filament go had it happen when the light was being turned on. That is a fact - the surge on turning on a light that is close to the end of its life will normally be the cause of the filament blowing and therefore the surge does have an effect on when the bulb finally blows. I never suggested that turning it on and off more frequently prior to that point has any effect on its remaining life, just that a bulb always fails in the end due to that surge. Presumably a bulb that is on constantly will eventually fail but since there are very few of those there is little opportunity to actually be there when one does in order to see it.

Obviously a light that is off is going to last a lot longer than one that is on. The surge that finally destroys the filament most likely comes within seconds of when the filament would have failed anyway without the surge.

The choice of Edison bulbs or LEDs prolongs the time before failure because they last longer than regular bulbs anyway and so take longer to reach the point toward the end of their life where a surge will finish them off.

The only connection any of this has to computers is that the same sort of situation will apply with many of the computer components where the slight surge on powering on a component that is already close to failing will probably be what finally finishes it off and the components will therefore most likely fail as they are being powered up.

As for the order in which components are likely to fail, I agree with Dan. The only components I have had actually fail due to age have been hard drives. I have had power supplies and motherboards fail but they were due to other issues that had nothing to do with how the computer was beng used or how long it was on for.

There are basically four states a computer can be in.

  1. Full power with all components operating during actual use
  2. Sleep mode where all the unneeded components turn off but there is still power to the memory.
  3. Power down/hibernate/standby mode where even the memory is powered down and only sufficient power is used to monitor for when the system needs to wake - front power button (always), wake on lan/modem/mouse/keyboard (if configured and the computer has not been off since last used)
  4. Completely off - either via an off switch on the power supply or at a poweboard or wall switch or by pulling the plug.

Any of the last three of these will extend the life of the hard drive.

45

Having thought about the causes of the failures I have had with computers I can say that surges have been responsible for most of them.

On one occasion the power to our neighbourhood went off for several hours due to someone running into a tree which then fell across the power lines. When it came back on the resultant surge blew the power supply in the one computer that had been shut down but not physically turned off.

On another occasion a lightning strike sent a surge through a number of devices on one of our power circuits as well as our phone system. This completely blew the motherboard and most of the other components inside a computer that was shut down but still connected to the power. The surge appears to have entered the computer via the lan. Tvs and other devices on the same electrical circuit that were in standby mode also blew. Other computers that were connected to the same lan but physically turned off were unaffected.

So from these examples I think it is reasonable to say that power surges can affect the life of your computer and that powering down a computer without physically turning it off leaves your computer more likely to suffer from such surges.

46

I never think it’s really necessary to unplug the computer after shut it down but you guys make me see more clear this problem…Does this unplugged make long life time to the computer ?

47

well, i power down usually an hour before i plug it out :confused2 - seems to be ok and ive never had a problem, i used to always do the defrag, but can you over do this really? some have told me it can be bad if u do it often :shifty:

48

If you’re running Windows 7, you never have to defrag, as the OS does it automatically in the background when it detects you’re not using the computer for a while. Defrag shows all of my disks at 0% fragmentation though I’ve never once run a manual defrag.

49

Funnily enough (or rather not) the component that dies first is my graphics card. On my current Mac Pro I’ve installed my THIRD damn graphic card in a little over a year and a half! They die pretty much after six months.

Nothing else has ever died on me, except one Raptor HD after a few months.

Were it not a Mac device, I’d have said “screw ATI” or rather that particular product, but alas, there are practically no other graphic cards available for my 2007 system. I’ve seen a modded version offered for my system in the US. I guess I’ll try that and hope that this graphic card will last at least a year. These are the times where I’d love to use a PC instead.

50

vista here Dan, I’m thinkin of going to Windows 7 though but I dont know enough about it yet - havent had the time; but that feature appeals to me for sure - where do u find out ur percentage without sounding too much of a newbie!?

I know for sure my laptop battery is on its way out, stupid thing - im not sure whether it has something to do with the latest update with windows BUT my battery indicator icon is showing at like “25% available plugged in - charging” for ages - this has never happened before - it doesnt shut down or anything so I’m assuming its just something finicky, im not worried about it but I guess i did the damage with constantly leaving the laptop charging anytime i use it, thats the only time anything has ever been troublesome to me; well and now that I think of it… on my old base unit - it stopped working too kept getting the BSOD, turned out the processor was just dusty, damn AMD’s and their overheating but a little dusting did the trick, im also thinking of going back to the base unit days :weee:

51

Most defects are manufacturing problems. Defects that may appear months or years later.

Other problems - ie felgall’s surge damage - are directly traceable to failure to install effective protectors. For example, if using a protector adjacent to the computer, then surge damage has been made easier. Those protectors do not even claim (see numeric specs) to provide necessary protection.

Your telco’s computer is connected to overhead wires all over town. It is threatened by about 100 surges with each thunderstorm. And must never fail. How often has your town been without phone service for four days while they replace that computer?

Informed homeowners do what every telco does everywhere in the world. They don’t waste massive sums on those ineffective plug-in protectors. They spend many times less money to install a protector where the wire enters the building and within feet of the single point earth ground. A protector without that short connection to earth is ineffective.

To make the protector even better, your telco wants its protector up to 50 meters distant from electronics. Effective protection is made even better when separation exists between the protector and electronics.

Effective ‘whole house’ protector (properly earthed) means a direct lightning strike to incoming wires is unknown. Even the protector must conduct that direct lightning strike and not fail. Nothing new. This was understood and routinely installed even 100 years ago.

Which demonstrates how many know without first learning facts. Which demonstrates how so many (a majority) use retail salesmen as a replacement for science and knowledge.

Ham radio operators in the early twentieth century would disconnect their antenna. Even put the antenna lead into a mason jar. And still suffer damage. Disconnecting is ineffective protection. Damage stopped when the antenna lead was earthed. That is what the ‘whole house’ protector does. That is why your telco provides telephone service during every thunderstorm and must not suffer damage. Please learn the science. Please dispose of myths that promote those ineffective plug-in protectors.

Whereas manufacturing defects are the most common reason for computer failure; still every informed homeowner upgrades earthing to make one ‘whole house’ protector so effective.

Even unplugging will never be as effective or as reliable as one ‘whole house’ protector.

52

OK First off, I personally leave the computer on… though if I’m AFK for a long period of time I’ll hit sleep mode, I have surge protectors (etc) to keep the lightning storms away and it’s fine. I usually do it so people can leave me a message on instant messenger (for the ones offline messaging isn’t available). :slight_smile:

As for the life of a PC, there is enough scientific proof that leaving your machine running rather than cold booting post shut down will extend the life of your PC. The main reason why this occurs is not down to heating or power, it’s down to those delicate devices we call hard disks, because conventional hard disks are finely tuned when cold booting the hard disk is put under a temporary state of increased pressure (both in usage and the initial power hit to the magnetic bits and pieces inside). Arguably how much longer the PC will last is highly debatable and to be honest most people will upgrade their machines before the thing is likely to fail, but I can say for sure that I’ve seen several studies by hard disk manufacturers and well known technology experts (like Steve Gibson) who have concluded that cold booting a machine to a REALLY tiny extent increases the chances of failures (barely noticeable extent) and the concept of keeping your machine running (or in an active state - like standy / sleep with the HDD left on) has been proven effective. I should also point out that hibernate interestingly enough increased the chances of error even more than shutting down on the basis of the memory dump to the HDD being recalled on top of the initial boot stresses. :slight_smile:

53

I would shut down the PC completely.
This will save the power too.

54

I have had bad experiences hibernating my PC…it winds up just shutting it down completely. I do not know if it was just a bug, but I didn’t really see the whole point of hibernating it.

55

Surge protectors don’t protect against lightning strikes. The surge from a lightning strike is generally many thousands of times the maximum rating of the surge protection.

56

My power outlets have ground and they are working properly :slight_smile: (like windows said : this device is working properly, it sounds strange for my latin language :slight_smile: )and like Alex said I don’t like cold booting, I think the simple shut down is enough…

57

Which is why plug-in proptectors are a scam promoted by myths purveyors. Which is why a $3 power strip with some ten cent protector parts and expensive paint sells to the naive for $150 with a Monster Cable label.

Reality demosntrates how many routinely ignore reality for the same reason why so many know power cyclng is destructive.

How does your telco’s computer - connected to overhead wires all over town - that suffers about 100 surges with each thunderstorm - how does that computer not suffer surge damage from direct lightnings strikes? One is supposed to learn 100 year old technology before posting. Telcos install less expensive protectors, short to earth ground, and distant from electronics. Then telcos all over the world suffer no damage. Therefore towns all over the world have phone service all four days after every thunderstorm.

Protection from direct lightning strikes is routine if 1) you do not listen to myth purveyors, 2) do not waste tens or 100 times more money on plug-in protectors, and 3) properly earth only one ‘whole house’ protector. That knowledge is well over 100 years old. And only understood by a minority who learn facts rather then be trained by hearsay from retail salesmen.

Too often, surges get blamed for failures traceable to manufacturing defects - when only observation and speculation results in a conclusion.

58

That may be true but this house has cut-off switches (fuse protectors) for when lightning hits the house and when it happens none of the equipment gets damaged. My surge protector has an external battery to assist my laptop so when everything goes out it kicks in the power when I need it to ensure I can carry on working. :slight_smile:

59

I lost a computer to a lightning strike, while plugged in to a surge protector.

Partly because this was back when dialup was the only internet access available, and the surge went through both the electric and the phone lines.

A TV, refrigerator, and all the phones were dead after it happened. The modem in the computer was fried.

60

None of that is surge protection. How does the millimeter separation in that ‘cut-off switches’ stop what even three miles of sky could not? How does a switch that takes tens of milliseconds to open stop a surge that is done, destructively, in microseconds?

I constantly complain about people who ‘know’ but never bother to first learn the numbers. What is the voltage on that cut-off switch? Hundreds of volts? That means an open switch continues to conduct electricity if the voltage is thousands of volts. More numbers that say a cut-off switch is an example of ‘urban myth’.

That external battery does what? From a first semester electrical course, that battery is nothing more than a conductor - electrically equivalent to a wire - to surge currents. That battery does nothing for protection. Even manufacturer specs numbers do not claim that protection. But again, a conclusion complete vanquished once we demand what are always required - technical numbers.

Surge protectors are always installed for surges that can overwhelm existing appliance protection. Those rare surges are lightning. Why do telcos all over the world suffer about 100 surges with each lightning storm - without computer failure? Still provide telephone service all four days afterwards? Because telcos install surge protectors that actually do protect from direct lightning strikes. That means telcos do not waste money on ineffective and obscene profit devices from APC, Belkin, Tripplite, Cyberpower, or Monster Cable. Instead telcos spend tens or 100 times less money for protector that harmlessly earth direct lightning strikes.

Telcos have been doing for 100 years what so many today deny. Deny because retails salesmen, for some stange reason, are electrical experts. Deny if a consumer never learns protection is always about where energy dissipates. A protector is only as effective as its earth ground - which plug-in protectors do not connect to and avoid all discussion of. Protection is a discussion of what provides protection - single point earth ground. Better for profit margins is to encourage myths that promote ineffective and undersized power strip protectors.

Don Grossman demonstrates what has been well undestood for decades. Dr Martzloff describes Don’s damage in his 1994 IEEE paper. Don’s plug-in protector, also called ‘point of connection’ device, may earth surges destructively through nearby appliances. Most believe only what they were told to believe. Forget to ask simple and damning questions. This Dr Martzloff’s first conclusion is not what most learn:
> Conclusion:
> 1) Quantitative measurements in the Upside-Down house clearly show objectionable
> difference in reference voltages. These occur even when or perhaps because,
> surge protective devices are present at the point of connection of appliances.